Can I recover the audio in my project directory?

I was recording a podcast for about an hour on my laptop, when suddenly, Audacity just disappeared. Long story short, my laptop had run out of hard drive space, and Audacity must have crashed when it ran out of storage or buffer.

However, my project directory has 1.7 GB of data in it, so I think there’s a good chance Audacity auto saved most of it.

But when I open the project file, there are no tracks visible, it opens as if it were an empty project.

I’ve Googled for how to recover data, and I’ve read that Audacity 2.1.2, which is what I’m using, has “automatic crash recover”. But, if that option does not appear automatically on the next time one opens Audacity, there seems to be no way to invoke it manually.

How can I recover all the audio that is sitting there in my project directory?

Unfortunately, running out of disk space is one of the worst things that can happen to Audacity. It not only trashes the “project file” but trashes the crash recovery data too (when Audacity tries to update the project structure information, there is no space to write into, and then the system shuts down.) The developers are working on improving this, but it’s tricky.

“Audacity” can probably not recover the data, but as your project was a new unedited recording, it may be possible to piece the data fragments back together. Each data block represents about 6 seconds of audio, and each block is time stamped. Stereo tracks are tricky to deal with, but depending on your file system, the time stamps may have fine enough granularity to indicate which files belong to which channel (I believe that the right channel block is written after the left channel block).

The project is too big to put together manually (it would be quicker to re-record), but it may be possible to bulk-rename the files based on their time stamp, then piece them back together into a single audio file. If your recording was mono, then I think the SoX script (at the bottom of the linked page) is probably the best chance, but I’ve never tried this myself. See:

Thank you for replying.

There were two audio tracks, each was mono.

I see in the project directory there are 7 directories, each containing about 256 items each. Are you saying it’s possible I might be able to append them and reconstruct an audio track? I could, with some help, possibly create a batch script to do that.

It’s hard to tell just by looking at them, but they seem to have a numbering system to them. Can I rely on them being sequentially numbered so that I could write a batch script that appends them one after the other by alphanumeric order of their name?

They all have the file extension “au”. Can I create one big “au” file out of all of them and then just open that in Audacity? Is “au” an audio format, like MP3 or WAV, or does it require special handling?

These questions are mostly covered by the page in the manual that I linked to. Please read this page carefully:

Yes, if the tracks have not been edited.

No they are not sequentially numbered.
You will need to rename them sequentially based on their file creation / modification time.

Yes. It is based on the AU format originally introduced by Sun Microsystems.

You need to strip off the file headers before concatenating the files.